silesr2015 004.

A Recorded Text Survey of

the Mfumte Speech Varieties

and Rapid Appraisal Survey of

Dzodinka

Varieties and Rapid Appraisal Survey of Dzodinka

Greg L. McLean

SIL International

®

2015

SIL Electronic Survey Report 2015-004, February 2015 © 2015 SIL International®

The purpose of this survey was to determine the language development needs of the Mfumte people. The Mfumte ‘people under the palms’ live in the North West Region of Cameroon and in Taraba State, Nigeria. A Rapid Appraisal (RA) survey and a Recorded Text Testing (RTT) survey were carried out as the continuation of an RA survey completed in 1994. There were two goals in this present survey: (1) to determine the potential extensibility of literature based on the Kwaja, Koffa, and Lus speech varieties to the surrounding communities and (2) to understand the language use patterns, vitality, attitudes, and inter-intelligibility in the neighbourhoods of Adere. Based on the results of the RA and RTT surveys, it is recommended that language development work be started in three speech varieties: Adere (Dzodinka [add]), Koffa, and Kwaja [kdz]. (Lus is already under development.) By identifying the language needs in the Mfumte area, work can proceed on creating orthographies, dictionaries, and other written materials.

iii Abstract

1 Introduction

2 Background information

2.1 People and languages 2.1.1 Mfumte 2.1.2 Kwaja 2.1.3 Ndaktup 2.1.4 Dzodinka 2.2 Neighbouring languages 2.3 Location

2.4 Previous research 2.5 Demographic situation

2.6 Languages of wider communication 2.7 Historical background

3 Goals and research questions

4 Methodology

4.1 Rapid Appraisal questionnaire 4.2 Rapid Appraisal wordlist 4.3 Recorded Text Testing purpose 4.4 Recorded Text Testing procedure 4.5 Post-Recorded Text Testing interviews 4.6 Recorded Text Testing results interpretation

5 RTT survey results

5.1 Results and interpretation of Kwaja story 5.2 Results and interpretation of Koffa story

5.2.1 Manang 5.2.2 Mbah

5.3 Results and interpretation of Lus story 5.4 Results of post-RTT interviews

6 RA survey results of Dzodinka

6.1 Dialectology

6.1.1 Adere quarters

6.1.2 Names and origins of the people 6.1.3 Lexicostatistics

6.2 Language vitality 6.2.1 Multilingualism

6.2.2 Migration and intermarriage 6.2.3 Language use

6.2.4 Language maintenance and shift 6.3 Language attitudes

6.4 Development potential

6.4.1 Homogeneity of the linguistic community (social cohesion) 6.4.2 Openness of the community to change

6.4.3 Presence of local middle-aged leadership 6.4.4 Conclusion on development potential 6.5 Summary of results

7 Conclusions and recommendations

7.1 Rapid Appraisal survey of Adere quarters 7.2 Recorded Text Testing survey

7.4 Recommendations for language development

7.5 Recommended modifications to Ethnologue and ALCAM

Appendix A: RTT Texts in Koffa, Lus, Kwaja, and Pidgin Appendix B: RTT Results

Appendix C: ALCAM 120-item Wordlist for Four Adere Quarters

Appendix D: Rapid Appraisal Group Questionnaire Appendix E: Religious Leader Questionnaire References

1

A Rapid Appraisal survey and a Recorded Text Testing survey were carried out in the Mfumte area in the North West Region of Cameroon between March and August, 2010. Rapid Appraisal (RA) attempts to carry out a quick assessment of language use patterns, vitality, attitudes, and inter-intelligibility between different language communities (cf. Bergman 1991; Stalder 1996). The research tools used in RA are questionnaires and wordlist elicitation. Recorded Text Testing (RTT) attempts to measure the degree of understanding that one speech community has with another using a recorded story (cf. Casad 1974). RTT also suggests whether a community’s degree of understanding is due to inherent intelligibility or

acquired intelligibility.

There are sixteen Mfumte villages: fourteen villages are in Nwa subdivision, Donga-Mantung division, North West Region, Cameroon, and two villages are located in Taraba State, Nigeria. The survey team consisted of Greg McLean (SIL), Cameron Hamm (SIL), William Nwufa, and Novester Nweshie. Cameron Hamm, who has had extensive experience in language survey work, was able to advise the team members on survey methodology, especially in the initial weeks of the survey.

The research team wishes to thank all government, church, and traditional leaders for their

assistance and hospitality. Team members also wish to thank the local people who gave of their time and knowledge. The author also wishes to thank the following people for their invaluable comments on this report: Robert Hedinger, Randy Lebold, Larry Seguin, and Cameron Hamm. Any shortcomings in this report are the responsibility of the author.

The body of this report consists of seven sections: (1) introduction, (2) background information, (3) goals and research questions, (4) methodology, (5) RTT results, (6) RA results of Dzodinka, and (7) conclusions and recommendations. The introduction outlines the research that took place in the Mfumte area. The second section gives some background information on the Mfumte people, their linguistic context, and previous research. The third section defines the purpose of the research and presents the questions that the researchers sought to answer. The fourth section details the methods and principles underlying Rapid Appraisal survey and Recorded Text Testing. The fifth and sixth sections present the results of the research. The seventh and final section states the conclusions of the present research and offers some recommendations for future language development.

2 Background information

The following subsections provide a brief overview of the languages and social context of the Mfumte people.

2.1 People and languages

The name Mfumte is used by outsiders to describe the people group. The Mfumte people also refer to themselves as Mfumte. It is a name borrowed from the Limbum language and means “people who live under the palms.” The term Mfumte is also used to refer to the language spoken by a subset of the Mfumte people. The Atlas Linguistique du Cameroun (ALCAM) lists two languages for the Mfumte people: Mfumte [905] and Dzodinka [904] (Dieu and Renaud 1983). ALCAM includes all the villages as speaking Mfumte,1 _{except Adere, where Dzodinka is spoken.}

In Ethnologue (Lewis et al. 2014) there are four entries that describe the languages of the Mfumte people. The first is Mfumte [ISO 639-3: nfu] which groups together the central and western Mfumte

villages: Lus, Kom, Mballa, Bang, Koffa, Jui, Mbat, Manang, Mbibji, and Mbah.2 _{The other three entries}

in Ethnologue that describe the Mfumte people are Kwaja [ISO 639-3: kdz] which refers to the language spoken in the village of Kwaja; Ndaktup [ISO 639-3: ncp] which refers to the two dialects spoken in Bitui and Ncha; and lastly, Dzodinka [ISO 639-3: add] which describes the language spoken in the different quarters of Adere. During this survey, it was revealed that the people of Adere prefer to call their language Lidzonka. In this report the author maintains the use of the term Dzodinka (not Lidzonka) for the sake of consistency with Ethnologue and ALCAM. However, it is recommended that the name be changed to Lidzonka in the future editions of these publications (see section 7.5).

Unless otherwise noted, in this report the term Mfumte refers to the people group and not a specific language.

2.1.1 Mfumte

The classification of Mfumte found in Ethnologue is Niger-Congo, Atlantic-Congo, Volta-Congo, Benue-Congo, Bantoid, Southern, Wide Grassfields, Narrow Grassfields, Mbam-Nkam, Nkambe, Mfumte. ALCAM

classifies Mfumte [905] as Niger-Kordofan, Niger-Congo, Bénoué-Congo, Bantoïde, Bantou, Grassfield, Est-Grassfield, Nord, Mfumte.

2.1.2 Kwaja

The classification of Kwaja found in Ethnologue is Niger-Congo, Atlantic-Congo, Volta-Congo, Benue-Congo, Bantoid, Southern, Wide Grassfields, Narrow Grassfields, Mbam-Nkam, Nkambe, Kwaja.

2.1.3 Ndaktup

The classification of Ndaktup in Ethnologue is Niger-Congo, Atlantic-Congo, Volta-Congo, Benue-Congo, Bantoid, Southern, Wide Grassfields, Narrow Grassfields, Mbam-Nkam, Nkambe, Ndaktup.

2.1.4 Dzodinka

The classification of Dzodinka in Ethnologue is Niger-Congo, Atlantic-Congo, Volta-Congo, Benue-Congo, Bantoid, Southern, Wide Grassfields, Narrow Grassfields, Mbam-Nkam, Nkambe, Dzodinka. ALCAM

classifies Dzodinka [904] as Niger-Kordofan, Niger-Congo, Bénoué-Congo, Bantoïde, Bantou, Grassfield, Est-Grassfield, Nord, Dzodinka.

2.2 Neighbouring languages

In Cameroon there are three languages that directly surround the Mfumte area. Two of these are Grassfields languages: Yamba [906]/[yam] and Limbum [903]/[lim]. To the northwest, there is a Jukunoid language called Mbembe [700]/[nza]. In Nigeria, to the north and northeast, there are up to nine languages bordering the Mfumte area, according to Ethnologue. Two of the languages are Tivoid: Abon [abo] and Batu [btu]. Three are Mambiloid: Ndoola [ndr], Ndunda [nuh], and Mambila, Nigeria [mzk].

A group of four other languages is concentrated in a small area to the west of the Mfumte area: Áncá [acb],3 _{Fum [fum],}4 _{Kwak [kwq],}5 _{and Viti [vit]. Áncá and Kwak are both listed as Niger-Congo,}

Atlantic-Congo, Volta-Congo, Benue-Congo, Bantoid, Unclassified. Fum is classified as Niger-Congo, Atlantic-Congo, Volta-Congo, Benue-Congo, Bantoid, Southern, Wide Grassfields, Narrow Grassfields. Viti is listed as Niger-Congo, Atlantic-Congo, Volta-Congo, Benue-Congo, Bantoid, Southern, Wide

Grassfields, Narrow Grassfields, Unclassified (Lewis et al. 2014).

2.3 Location

The Mfumte are primarily located in Nwa subdivision, Donga-Mantung division, North West Region of the Republic of Cameroon.

Map 1. Mfumte area

3_{Áncá [acb] “may be the same as Manta [myg]” (Lewis et al. 2014).}

4_{Fum may be the same as one of the varieties found in the Mfumte area (Lewis et al. 2014).} 5_{Kwak may be the same as Yamba (Lewis et al. 2014).}

2.4 Previous research

Previous academic research in the Mfumte area dates back to the early 1980s. Voorhoeve (1980) wrote a paper on noun classes in Adere. A 120-item wordlist was transcribed in the Dzodinka language by Voorhoeve and Leroy in connection with the ALCAM project (Dieu and Renaud 1983). Anthropologist Viviane Baeke has written an ethnography that focuses on the cultural traits found in Lus, one of the Mfumte villages where she lived off and on from February 1980 to August 1983 (Baeke 2004).

Grant et al. completed a Rapid Appraisal survey describing the language situation in the Mfumte area (1994). The RA survey of 19946 _{suggested that the fourteen Mfumte villages be divided into four}

language groups (Grant et al. 1994:7–19). This division into four groups was based on reported comprehension and was confirmed by lexicostatistic analysis. Linguistic differences and sociolinguistic attitudes between the four groups were noted. Grant et al. say, “…due to its linguistic distinctiveness Dzodinka should be codified quite apart from the other Mfumte villages.” They also note that “some villages (Jui, Manang, Kom, Koffa, Bang) expressed real reservations towards Lus being used as the written Mfumte language.” Lexicostatistical analysis revealed that the lexical similarity between the wordlists of Bitui, Ncha, and Kwaja was quite low when compared to the village of Lus—with

percentages of 35, 29, and 23, respectively.7 _{The threshold figure for grouping speech varieties together}

was 60 percent. Grant et al. said that respondents in Ncha and Manang reported that the vast majority of the people in their villages had no comprehension of the speech variety spoken in Lus (1994:7).

Lexicostatistical analysis revealed that the lexical similarity between wordlists of Bitui, Ncha, and Kwaja was quite low when compared to the village of Koffa8_{—with percentages of 36, 29, and 28, respectively}

(1994:11). The linguistic differences between some groups and the negative sociolinguistic attitudes between other groups suggest that none of the four groups should be combined.

The present survey is written approximately two decades after the original RA survey. Since the original survey, there have been no major changes in sociolinguistic patterns in the Mfumte area. There has been no major influx of foreigners or inter-village migration that could cause a major shift in sociolinguistic patterns. Although there are more motorcycles in use now and the network of roads and paths has improved slightly, most villages are as secluded as they were in 1994. The negative attitudes expressed by some respondents in the earlier survey towards the use of the Lus speech variety have also been expressed to the present author. In the absence of any major change in sociolinguistic patterns, the author affirms the four village groupings as proposed in Grant et al.

Grant et al. grouped the villages of Ncha, Bitui, and Kwaja together on the basis of reported

comprehension, and they said that lexicostatistic analysis was in agreement with this decision. They also reported, “Ncha and Bitui group together and claim to easily understand Kwaja; on the other hand, Kwaja does not perceive as close a relationship with these other two villages.”9 _{Despite Ncha and Bitui}

reporting good comprehension of Kwaja, Grant et al. recommended that ALCAM create separate entries for Ndaktup (the language spoken in Ncha and Bitui) and Kwaja. As will be seen, the present RTT survey suggests that Ncha and Bitui have good comprehension of Kwaja. Consequently, it will be suggested that the development of one language (i.e. Kwaja) will be adequate to meet the linguistic needs of the surrounding villages.

6_{The two additional Mfumte villages of Antere and Inkiri (located in Nigeria) included in the present survey were} not included in the 1994 RA survey.

7_{The margin of error was ± 3.6–9.8 percent.}

8 _{Koffa was used for comparison since it is the biggest village in its group (see table 1) and also because the speech} variety spoken there is the most widely understood by neighbouring villages.

Table 1. Village groupings in the 1994 RA survey

Group 1 Group 2 Group 3 Group 4

Adere Ncha Lus Kom

Bitui Inkiria _Mballa

Kwaja Bang

Anterea _Koffa

Jui Mbat Manang Mbibji Mbah a_{In the current survey, these Nigerian villages were added to Groups 2 and 3.}

As a starting point for the RTT survey, the current survey used the suggested four village groupings of the 1994 survey. The decision to include Antere in group 2 and Inkiri in group 3 (see table 1) was based on the informal questioning of the local people. The author was also able to confirm the close similarity between the languages spoken in Inkiri and Lus by his ability to communicate with people from Inkiri.10

As part of the present research, an RA survey was administered in several quarters of Adere. Although Adere (Dzodinka) was included in the 1994 RA survey, there remained a need to examine the inter-intelligibility between the speakers of the diffuse quarters. In preference to an RTT survey, an RA survey of the Adere quarters would provide a cursory understanding of language use patterns, vitality, attitudes, and inter-intelligibility.

The most recent linguistic field research in Lus was done by Julius Eyoh working under BASAL (Basic Standardization of All African Languages)—a program of the National Association of Cameroonian Language Committees. A phonology (Eyoh 2008), tentative orthography (Eyoh and Hedinger 2008), alphabet chart, wordlist, primer, and other materials were produced based on the dialect spoken in Lus.

2.5 Demographic situation

Ethnologue reports that there are 24,700 people who speak Mfumte; 2,980 people who speak Kwaja, 2,980 people who speak Ndaktup; and 2,600 who speak Dzodinka. The author believes that the 1982 figure of 24,700 for Mfumte speakers was an overestimate. The author understands this figure to represent the population of the ethnic group and not the number of speakers of a specific language. The local people identify themselves as Mfumte but have not historically used the term Mfumte to describe their language. Ideally, this would be reflected in Ethnologue and ALCAM by removing the entry of Mfumte as a language name. Based on the four groups found in table 1, the author posits new11 population estimates for 2014 in table 2.

10_{At the time of this survey, the author had spent over two years learning the language spoken in the village of Lus.} 11_{The author estimates new population figures because, except for Adere (Dzodinka), the village groupings in} table 1 do not correspond with the language entries found in Ethnologue.

Table 2. Population estimates for the Mfumte people

Group Villages 2014 population

1 Adere (Dzodinka) 3,574a

2 Ncha, Bitui, Kwaja, and Antere 5,500

3 Lus and Inkiri 6,500

4 Kom, Mballa, Bang, Koffa, Jui, Mbat, _{Manang, Mbepji (Mbibji), and Mbah 12,000}

a_{Ethnologue lists Dzodinka (Adere) as having 2,600 speakers (this estimate is from the}

year 2000). The 2014 estimate is arrived at by applying a growth rate of 2.3% per annum (World Bank figure) to the original estimate.

2.6 Languages of wider communication

The Mfumte people use Cameroon Pidgin [wes] as a language of wider communication, especially when they perceive that their own language will not be understood. English is the medium of instruction in all nursery, elementary, and secondary schools in the Mfumte area.

2.7 Historical background

The term Mfumte, a name given by the neighbouring Wimbum people, is an umbrella term for a people who historically came to the area from different places at different times. As they arrived, these people would have spoken the mother tongue of their places of origin, as well as any other learned languages. The name Mfumte, means “people under the palms”—an apt description of the multitude of palms that exist in the area. People from Lus trace their origins from many different villages: Lasin, Mbot, Njari, Adere, and Ntsɨ’. In these villages the languages that are currently spoken are Noone [nhu], Limbum [lim], Mbembe [nza], Dzodinka [add], and Lamnsoʼ [nso], respectively. That some Lus people reported that they were originally from the Ndu area is contrary to that which was reported by Baeke, “…mais aucun lignage wuli ne se dit originaire du pays des Nsungli”12 _{(2004:25). Nsungli is a name that refers to}

the Wimbum people. Baeke also reported that the Lus people have a closer cultural relationship with Kwaja than with Wimbum, but linguistically, the Lus people are closer to the Wimbum people (2004:25). In Lower Bang and Bugo (quarters of Adere), the people reported that they originated from the Tikar area and also from Abong, Nigeria. The people in the Tikar area and Abong currently speak Tikar [tik] and Mbembe [nza], respectively.

3 Goals and research questions

The overarching purpose of this survey was to determine the language development needs of the Mfumte people. As such, there were two goals: (1) to determine the potential extensibility of literature based on the Kwaja, Koffa, and Lus speech varieties to the surrounding communities and (2) to understand the language use patterns, vitality, attitudes, and inter-intelligibility in the neighbourhoods (also called quarters) of the village of Adere. This understanding helps to determine the potential extensibility of literature based on the Bugo speech variety to the surrounding quarters of Adere (which are Bugo, Bankem, Bangang, Nchukun, Lower Bang, and Upper Bang).

In order to realize the first goal, the researchers gathered and analysed information according to the Recorded Text Testing method. The decision to use the speech varieties spoken in Kwaja, Koffa, and Lus

was based on the suggested village groupings and ease of comprehension described in Grant et al. (1994). The researchers sought to answer three main questions:

1. Could literature based on the Kwaja speech variety be used in the villages of Bitui, Ncha, and Antere?

2. Could literature based on the Koffa speech variety be used in the surrounding villages of Manang, Mbat, Jui, Bang, Mballa, Kom, Mbah and Mbepji?

3. Could literature based on the Lus speech variety be used in the village of Inkiri?

In order to realize the second goal, the researchers gathered and analysed information according to the Rapid Appraisal method. This method includes research into sociolinguistics, using Rapid Appraisal group questionnaires, and lexicostatistics. The researchers wanted to answer the following questions:

1. Do the speech varieties of the quarters of Adere show a high level of lexical similarity based on wordlists?

2. What is the potential for language development based on language vitality (see section 6.2), language attitudes (see section 6.3), social cohesion, attitudes towards community development, and demographics (see section 6.4)?

3. Could literature in Bugo (the quarter where the chief resides) be used in the surrounding quarters of Adere?

The desired outcome for this research was to gain insight into the language development needs in the Mfumte area. By identifying language needs, work can proceed on creating orthographies,

dictionaries, and other written materials. The desired impact is for the Mfumte people to be able to read and write in a language that is, or closely represents, their mother tongue.

4 Methodology

The research tools used in RA are questionnaires and wordlist elicitation. The questionnaires help to address the question, “What are the language use patterns, vitality, attitudes, and inter-intelligibility in a given region?” Wordlist elicitation and subsequent lexicostatistical analysis help to indicate which speech varieties can be grouped together.

The tools of RTT are comprehension testing and post-RTT interviews. RTT attempts to measure the degree of understanding that one speech community has with another, using a recorded story. It attempts to answer the question, “Is literature based on one speech variety extensible to other

communities?” RTT also aids in answering the question, “Are peoples’ comprehension levels a result of inherent intelligibility or acquired intelligibility?” In the present survey, the post-RTT interviews helped to answer the question, “What are the attitudes of the people towards the speech varieties being tested?”

The following subsections describe the tools used in RA and RTT.

4.1 Rapid Appraisal questionnaire

The Rapid Appraisal method helps researchers to understand the sociolinguistic conditions existing in language communities. It makes use of a group questionnaire (see appendix D). The information gathered from the questionnaires reveals the level of multilingualism that exists among the people, the patterns of language use, and the degree of contact between local people (Bergman 1991; Stalder 1996). The group questionnaire also seeks to examine the attitudes (especially those of traditional, political, and educational leaders in the community) concerning language development. Responses elicited by the questionnaire are also useful in evaluating language vitality and viability. In this survey one RA group questionnaire was administered in each Adere quarter visited by the survey team. The groups consisted of around fifteen to twenty-five people and included a cross-section of indigenous residents. In a group interview it is possible that the younger people or females could yield to the opinions of older males. In this survey however, the research team did not perceive this to have taken place.

4.2 Rapid Appraisal wordlist

The author collected 120-item wordlists from mother tongue speakers in four Adere quarters (see

appendix C), namely, Upper Bang, Bugo, Nchukun, and Bangang. Lexicostatistical analysis can reveal the level of lexical similarity between two or more speech varieties. This is one method used to determine if specific speech varieties can be grouped together. Wordlists were gathered from residents who had grown up in the quarters and who had spent only limited amounts of time (up to two years) outside the community. An attempt was made to obtain wordlists from people whose parents were also from the same quarter. The percentage of apparent cognates was then calculated by comparing the lexical pairs found in the wordlists.

The present survey employed the inspection method of comparing lexical items. The inspection method uses a two-point scale: lexical pairs are assigned zero points if they are dissimilar and one point if they are similar.

4.3 Recorded Text Testing purpose

Recorded Text Testing attempts to measure the degree of understanding (intelligibility) that one speech community has with another. RTT also suggests whether a community’s degree of understanding is due to inherent intelligibility or acquired intelligibility. A high rate of understanding would show that literature in one speech variety could likely be used by another speech variety.13 _{That said, RTT only}

tests the understanding of a simple story; therefore, the results of RTT testing should not be regarded as infallible.

4.4 Recorded Text Testing procedure

RTT consists of recording a short (two to three minutes) text of a speaker’s personal experience of an event that is not generally known and then testing the language comprehension of others with it. The story is recorded, and then questions about the story are created and translated into the speech variety of those who will be tested. After the creation and before the translation of the questions, other speakers (with the same mother tongue as the storyteller) are tested—this is known as the “hometown test.” In the “hometown test,” both the story and the questions are in the mother tongue of the test taker. Questions with recurring incorrect answers in the “hometown test” are removed, and the remaining questions are translated into another speech variety. The removal of bad questions allows the test results to be influenced by the comprehension of the story rather than by some other variable(s).

In order to remove participants who are not good test takers, a different story referred to as a sample test (Casad 1974) with accompanying questions is normally given before the main

comprehension test. Both the story and the questions of the sample test are normally in the mother tongue of the person taking the test. In this RTT survey the screening tests were given in Pidgin English. The use of a sample test in Pidgin saved the survey team a great deal of time by eliminating the need to elicit test stories in each of the twelve language varieties in question. By giving the sample test, the researchers were able to give the participants an opportunity “…to learn the mechanics of taking the test” (Casad 1974). The sample test also screens out participants who are too nervous, not mentally fit, not able to understand the testing process, etc. from taking the comprehension test. In this survey a participant who scored below seventy-two percent on the screening test did not proceed to take the comprehension test. Admittedly, using a test story in Pidgin English precluded a small number of participants who didn’t understand Pidgin well from proceeding to the comprehension testing. The screening of participants is intended to assure that language is the key factor influencing comprehension.

For comprehension testing, the participant is allowed to listen to the whole story through once. After that, the participant listens to the story again, this time in small sections which are each followed by a related question. After each section of the story is played and each question is asked, the recording is paused and the participant responds in his or her own mother tongue. A translator with the same mother tongue as the participant then translates the response into a language that the team members can understand. A team member records a number (3, 2, 1 or 0) for each question, based on the response of the participant/translator. Table 3 shows the meaning of each number.

Table 3. RTT scoring method Number

recorded

Response Points

3 Correct 1

2 Partially correct 0.5

1 Incorrect 0

0 Participant is distracted from hearing the recording or question.

Question is omitted from the results.

After the survey session, the recorded numbers are converted to points for the sake of analysis. A correct answer is given one point, a partially correct answer is given half a point, and an incorrect answer is given zero points. When the participant is distracted from hearing a particular section of the story and/or the related question, and is unable to answer the question adequately, a “0” is written down signifying that the question will be omitted from the results. This is done so that the results will not be negatively impacted because of something unrelated to the participant’s comprehension. Removing the question serves to differentiate between an incorrect answer that is caused by a lack of hearing well and an incorrect answer that is caused by a lack of comprehension.

Members of the survey team interview prospective participants before the testing begins. They collect data such as name, place of birth, age, sex, language spoken at home, mother’s language, father’s language, number of years lived outside of the area, and number of years lived in another village in the same area. The team members then use this data to select participants to be tested. The goal is to have at least ten participants complete the comprehension test. Ideally, there should be a good balance between male and female, young and old. Young is defined as those aged thirty-five years and under, and old is defined as those aged thirty-six years and over. The participant’s mother and father should come from the same village where the testing is taking place. The participant should not have spent a lot of time14

outside of the language area as this could negatively affect his or her test score. A participant who has lived outside the language area and has not spoken his mother tongue for many years can start to lose his ability to speak and comprehend his language. Additionally, the participant should not have spent a lot of time in the village of the test language as this will skew the test results. Because the survey attempts to determine inherent intelligibility, a participant who has lived in the village of the test language for a long time will likely score high because of acquired intelligibility and not inherent intelligibility. The team perceived that migration away from most Mfumte villages has been minimal. The small number of participants who have stayed outside of their villages have done so only for a short time (fewer than three years) in order to work or go to school.

Based on the suggested village groupings, the speech varieties spoken in Kwaja, Koffa, and Lus were chosen as test varieties. These groupings (see table 1) were determined by reported comprehension and confirmed by lexicostatistical analysis (Grant et al. 1994:11). The speech varieties spoken in Kwaja, Koffa, and Lus have the greatest numbers of speakers relative to the other villages in their respective

groups. According to reported comprehension, the speech varieties spoken in Kwaja, Koffa, and Lus are also the most readily understood by the largest number of surrounding villages in their respective groups (Grant et al. 1994:7–8).

4.5 Post-Recorded Text Testing interviews

Post-RTT interviews are normally conducted in conjunction with Recorded Text Testing, because they help to determine language attitudes toward the test variety. When writing this report, it became clear to the author that specific questions regarding the acceptance of the reference dialect would still need to be answered. The attitudes expressed towards neighbouring speech varieties in the 1994 RA survey are inadequate for answering the outstanding questions. For this reason, the author conducted post-RTT interviews several months after the comprehension testing portion of the survey. The timing of the post-RTT interviews is not viewed as significant.

The interviews consisted of two questions, and they took place informally during group meetings in the villages where a recorded story was tested. The principal question posed during these interviews was, “Would you be willing for your children to learn to read and write in X language (where X represented the tested variety)?” The answer to this question gave the research team an indication of the attitudes towards the speech variety being tested. Another question concerning the name of the language was asked, “Do you accept to call your language X (where X was Southern Mfumte, Eastern Mfumte, or Central Mfumte)?” The use of geographical terms to refer to the different language groups was more acceptable to the people than using language names that refer to a specific village.15 _{The name of the}

speech variety spoken in each Mfumte village (as pronounced in the mother tongue) typically follows the pattern, “language of (village name).” The two questions posed in the post-RTT interviews helped the researchers to understand the attitudes of people toward the tested variety and also to identify language names that would be acceptable to the people.

4.6 Recorded Text Testing results interpretation

There are two main things to examine when looking at the results of an individual RTT survey. The first is the average or mean test score of all those taking the test, and the second is the standard deviation from that average. An individual’s comprehension test score is calculated by adding up the points given for each of his or her responses. The total points are then divided by the number of questions (excluding any question(s) where the participant is distracted) and converted to a percentage. Each percentage score is then added up and divided by the number of participants. The resulting average shows, in a general way, the speech community’s comprehension of the language being tested. High average scores could imply that literature based on the language spoken in the story could be used in the community. Concerning the interpretation of intelligibility scores, J. Grimes (1995:22) states the following:

At threshold levels high enough to guarantee good communication from the central dialect to the periphery (usually 85 percent or above), it is reasonable to speak of the dialect cluster as a single LANGUAGE from the linguistic point of view. Speech varieties that come together at only 70 percent or below are too distinct to qualify as the same language. In between, 70 percent to 85 percent, is an area of MARGINAL intelligibility where some communication is satisfactory and some is not. The threshold depends on the risk associated with not communicating well; the final criteria are not purely linguistic.

Standard deviation measures the degree of difference between individual scores and the average. If the scores differ significantly from participant to participant, it indicates that the participants have varying levels of contact with the test language. That is, certain participants have learned the test

language (acquired intelligibility) through exposure whereas other participants have not. Conversely, when comprehension scores are consistently high among participants, it is probably due to the close linguistic similarities of the two speech varieties (inherent intelligibility). Likewise, when comprehension scores are consistently low among participants, it is probably due to the lack of linguistic similarities of the two speech varieties, which means that inherent intelligibility would not be possible. Regarding the level of standard deviation that is considered significant, B. Grimes (1987:50) states the following:

If the standard deviation of intelligibility scores from various speakers in one community is more than 15 percent, it probably indicates bilingualism. The figure of 15 percent is also a suggestion of a possible maximum difference there may be in true intelligibility.

When comprehension is above 85 percent and there is a standard deviation of less than 15 percent, literature in the test language could be used by the language communities who meet these criteria. This assumes that other sociolinguistic criteria, such as attitudes towards the test language, are favorable.

In contrast, when comprehension scores are between 70 and 85 percent, further examination is required. In this situation, standard deviation and community attitudes should be verified to predict if the community will have success in using literacy materials based on the test language.

When the scores are below 70 percent, the comprehension level of the test language is insufficient and literature based on it would not be able to be used in that language community.

5 RTT survey results

The RTT research results are organized by the three speech varieties in which stories were recorded and tested. The researchers sought to answer three main questions:

1. Could literature based on the Kwaja speech variety be used in the villages of Bitui, Ncha, and Antere?

2. Could literature based on the Koffa speech variety be used in the surrounding villages of Manang, Mbat, Jui, Bang, Mballa, Kom, Mbah, and Mbepji?

3. Could literature based on the Lus speech variety be used in the village of Inkiri?

5.1 Results and interpretation of Kwaja story

RTT testing of the Kwaja story took place in three villages located to the north of Kwaja—Bitui, Ncha, and Antere. Table 4 shows the mean score and standard deviation of each of the villages that were tested with the Kwaja story.

Table 4. RTT results using Kwaja story

Village Mean score (%) Standard deviation (%)

Bitui 98 4

Ncha 92 9

Antere 94 9

The villages of Bitui, Ncha, and Antere all scored above the 85 percent threshold for adequate comprehension. The three villages also scored below the threshold of a 15 percent standard deviation which suggests that their inherent intelligibility of Kwaja is sufficient. Therefore the scores suggest that literature based on the Kwaja speech variety could be used in the villages of Bitui, Ncha, and Antere.

5.2 Results and interpretation of Koffa story

RTT testing of the Koffa story took place in eight villages surrounding Koffa. Table 5 shows the mean score and standard deviation of each of the villages that were tested with the Koffa story.

Table 5. RTT results using Koffa story

Village Mean score (%) Standard deviation (%)

Manang 79 16.1

Mbat 98 4

Jui 97 5

Bang 92 10

Mballa 96 6

Kom 98 5

Mbah 82 17

Mbepji 96 9

The villages of Mbat, Jui, Bang, Mballa, Kom, and Mbepji all scored above the 85 percent threshold for adequate comprehension. These same villages also scored below the threshold of a 15 percent standard deviation which suggests that their inherent intelligibility of Koffa is sufficient. Therefore the scores suggest that literature based on the Koffa speech variety could be used in the villages of Mbat, Jui, Bang, Mballa, Kom, and Mbepji. In contrast, the mean scores for Manang and Mbah were below the threshold of 85 percent for adequate comprehension.

5.2.1 Manang

The comprehension scores of participants in Manang ranged from 56 percent to 100 percent. The mean score was 78.9 percent and the standard deviation was 16.1. The large variation in scores suggests that those who tested well probably had considerable contact with Koffa people or even other villages close to Koffa such as Mbat or Jui. Conversely, those who scored poorly probably had less contact with people from Koffa. The speech variety of Manang should be considered as a distant dialect. However, it is close enough to Koffa linguistically that it is worth including Manang in the Koffa language project. There is still a possibility that Manang could use Koffa literature since their comprehension scores are above the 70 percent threshold. If the attitudes of the Manang people are favorable to the use of Koffa materials, then that may overcome their possible linguistic difficulties of using Koffa.

Alternatively, if the Manang people have negative attitudes towards reading and writing in Koffa, further testing could be done to see if they could use literacy materials in the Limbum language. Referring to the Mfumte area, Grant et al. (1994:13) say, “As regards standardisation, as with Yamba, Limbum is limited in its usefulness as a language for written use, although materials could possibly be introduced in some Mfumte villages, most notably Mbepji, Mbat, and Manang.”

5.2.2 Mbah

The comprehension scores of participants in Mbah ranged from 44 percent to 100 percent. The mean score was 81.82 percent and the standard deviation was 17. The large variation in scores suggests that those who tested well probably had some significant contact with Koffa people or even other villages close to Koffa. Conversely, those who scored poorly probably had less contact with people from Koffa. The older Mbah participants (thirty-six years of age and above) scored much higher on the test than the younger ones. The older participants had a mean score of 95.56 percent versus 70.37 percent for the younger participants. The standard deviations for the older and younger participants were 6 and 15, respectively. This suggests that by virtue of their age, older participants have had more time to learn the Koffa speech variety than the younger ones.

If the Mbah people favor the use of Koffa materials, it is recommended that they be a part of the Koffa language project. Since Mbah is related to Koffa linguistically (comprehension scores were above 70 percent), there is a possibility that Mbah could use Koffa literature despite the standard deviation

being over 15. It is recommended then, if and when Koffa is developed, testing with Koffa alphabet books and primers be done in Mbah.

Alternatively, if the Mbah people have negative attitudes towards reading and writing in Koffa, perhaps they could use literacy materials in the Limbum language. Grant says, “The Mbah leaders also claimed that Limbum is well understood even by Mbah children of 5 to 6 years old.” While there are some linguistic similarities between the speech of Mbah and Limbum, the Mbah people have learned Limbum (acquired intelligibility) mostly through contact with Wimbum people (Grant et al. 1994:13).

5.3 Results and interpretation of Lus story

RTT testing of the Lus story took place in one village, Inkiri, which is located to the north of Lus. Table 6 shows the mean score and standard deviation for the village of Inkiri which was tested with the Lus story.

Table 6. RTT results using Lus story

Village Mean score (%) Standard deviation (%)

Inkiri 97 4

The village of Inkiri scored above the 85 percent threshold for adequate comprehension. The village also scored well below the threshold of 15 percent standard deviation which suggests that the people’s inherent intelligibility of the Lus speech variety is sufficient. Therefore the scores suggest that literature based on the Lus speech variety could be used in the village of Inkiri.

5.4 Results of post-RTT interviews

Post-RTT interviews were conducted several months after the recorded text testing. The interviews helped to determine language attitudes toward the test variety. The interviews consisted of two

questions, and they took place informally during group meetings in the villages where a recorded story was tested. The principal question posed during these interviews was, “Would you be willing for your children to learn to read and write in X language (where X represented the tested variety)?” The second question concerning the name of the language was asked, “Do you accept to call your language X (where X was Southern Mfumte, Eastern Mfumte, or Central Mfumte)?” The interviews took place during church meetings, family meetings, meetings at the chief’s palace, etc. The responses to these two questions were both extremely positive, as can be seen in table 7.

Table 7. Post-RTT interview results

Language group

Percentage of respondents willing that their children learn to read and write in the _X_ language

Percentage of respondents willing that their language be called _X_ Mfumte Eastern Mfumte—

Bitui, Ncha, and Antere

100 100

Southern Mfumte— Manang, Mbat, Jui, Bang, Mballa, Kom, Mbah, and Mbepji.

100 95a

Central Mfumte—

Inkiri 100 100

a _{A few men from Manang said that they would prefer to call the language “Upper}

6 RA survey results of Dzodinka

The purpose of doing a Rapid Appraisal survey in the quarters of Adere (Bugo, Bankem, Bangang, Nchukun, Lower Bang, and Upper Bang) was to understand the language use patterns, vitality, attitudes, and inter-intelligibility. This understanding will help to answer the question, “Could literature in Bugo (the chief’s quarter) be used in the surrounding quarters of Adere?”

6.1 Dialectology

The following subsections compare the speech varieties found in the different quarters of Adere. Subsection 6.1.1, “Adere quarters” describes the reported language similarity and comprehension of some of the Adere quarters. The subsequent subsections deal with the names and origins of the people and lexicostatistics.

6.1.1 Adere quarters

Map 2. The quarters of Adere

The reported linguistic similarity and comprehension of speech varieties spoken in five Adere quarters are shown in table 8. There are three possible entries: A, B, C, and blank (--). The quarters in which the questionnaires were administered are listed in the first column. The quarters of which language

Table 8. Reported linguistic similarity and comprehension of Adere quarters

Bugo Bangang Lower

Bang

Upper Bang

Nchukun Bankem Amba

Bugo -- A A A A A A

Bangang A -- A A A A A

Lower Bang A A -- A A A A

Upper Bang A A A -- A A A

Nchuckun A A A A -- A A

Bankem --

Amba --

KEY

A: Respondents group the speech variety in question with their own – reported to be “same language” or “closest/most similar to ours,”

– immediate comprehension reported to exist, even among children 4-5 years old.

B: Reported as different, but with which there is a good comprehension, even among children 4–5 years old.

C: Understood only by older children (10+ years) or adults. Comprehension is variable within the population.

D: No comprehension, except by a few speakers.

Although representatives of the quarters of Bankem and Amba were not interviewed,

representatives of the other five quarters reported that their language was the same as that spoken in Bankem and Amba.

6.1.2 Names and origins of the people

The quarters of Adere where Rapid Appraisal questionnaires were administered were Bugo, Bangang, Lower Bang, Upper Bang, and Nchukun. The inhabitants of these quarters identified themselves as Wunka and identified their language as Lidzonka. People in Nchukun said that they originated from Bankem and before that, Lower Mbot, a Wimbum village northwest of Ndu. In Bangang the people did not identify their place of origin. In Lower Bang the people said they came from Bugo and before that, Tikari. In Bugo the people said they originated from Tikari and Abong, Nigeria. In Upper Bang people said they originated from Lower Bang.

6.1.3 Lexicostatistics

Table 9 shows the results of the comparative analysis of the ALCAM wordlists collected from four Adere quarters. Each figure represents the percentage of lexical similarity between two different quarters.

Table 9. Lexical similarity between Adere quarters Bugo

100 Bangang

100 100 Upper Bang

100 100 100 Nchuckun

This lexical data is consistent with the conclusions found from the RA group questionnaires. The lexical data is also well above the 60 percent level for grouping dialects together (cf. appendix C).

6.2 Language vitality

The following subsections describe the language vitality of Dzodinka. Language vitality is described in terms of multilingualism, migration and intermarriage; language use; and language maintenance and shift.

6.2.1 Multilingualism

Representatives of the Adere quarters reported that they have limited comprehension of the languages surrounding them. The languages and speech varieties described were the languages spoken in the villages of Lus, Kom, and Mbepji and also the languages of Mbembe and Limbum. All the interviewees, with the exception of those in Nchukun, reported that they switched to Pidgin English when conversing with the speakers of the previously mentioned languages and speech varieties. Nchukun reported that some can speak a little Limbum and also understand it to some extent by the age of fifteen. They also reported that after the age twenty, some people can speak and understand Mbembe. No one reported any comprehension of any Yamba dialects.

Those interviewed stated that some amongst them had a little comprehension of the surrounding languages and speech varieties after the age of twenty. When asked which of these speech varieties they understood most easily, Nchukun and Bangang said Mbembe; Lower Bang said Limbum; Bugo said Lus; and Upper Bang did not answer. The fact that most reported that they switch to Pidgin English when they meet someone from a neighbouring village suggests that their comprehension of neighbouring languages is not adequate. Given their generally limited comprehension of the language communities surrounding them, the Adere people are unlikely to use the literacy materials of these language communities. The relationship of Pidgin English to Dzodinka is further discussed below.

6.2.2 Migration and intermarriage

All of the Adere quarters reported that most people marry someone from one of the Adere quarters. Other than rules prohibiting incest, there are no rules governing who marries whom. All quarters except Bugo said that foreigners do not come and live in their quarters. Some young people leave the Adere quarters to go to bigger towns in Nigeria and Cameroon in search of work.

6.2.3 Language use

Dzodinka, the language of Adere, is being used in all domains of daily life. Dzodinka is the primary language used to communicate to other Adere people, whether in the market or in the fields. All the interviewees reported that the mother tongue is used at home, with friends, and on the farms. In these same domains, Pidgin English was reported to be used rarely. Dzodinka is also used at the local markets and at the health clinic. People reported that in public, i.e. announcements, local council meetings, and traditional religious ceremonies, the mother tongue is primarily used. In school, even though the

language of instruction is in English, Adere people reported that the mother tongue is sometimes used to explain concepts, and it is also used during recess time. Since reports indicate Dzodinka is being used widely in the various domains, it is reasonable to conclude that Dzodinka is vital and viable for the long term.

There are many different Christian denominations in Adere. The Cameroon Baptist Convention (CBC), the Presbyterian Church of Cameroon (PCC), the Roman Catholic Church, Deeper Life, and the Charisma Chapel are all present in Adere. In each of these churches Dzodinka is used in certain domains. Prayer can be made in the mother tongue, English, and Pidgin English. Sermons are generally given in English or Pidgin and then translated into the mother tongue. Scripture reading is done almost

exclusively in English. The exception is one church in Upper Bang that sometimes uses a Pidgin Bible for Scripture reading. Spiritual songs were reported to be sung in the mother tongue, English, and Pidgin. It was reported that there is a real need for translation, because large segments of the congregations (e.g.

the elderly, the youth, and those with limited education) are unable to understand English or Pidgin adequately. Information gathered through the group questionnaires showed that a significant percentage of people in each quarter are Christian (as high as 80 percent and as low as 50 percent). The remaining people hold to traditional beliefs and practices.

6.2.4 Language maintenance and shift

One way to understand if the use of the mother tongue is “shifting” or changing is to ask questions about how the younger generation use it. Respondents in the Adere quarters reported that their youth are not speaking another language more than the mother tongue. They also reported that the youth have a good attitude towards the use of their mother tongue. It was reported that the youth are mixing Pidgin English words into their mother tongue. When asked if this was good or bad, all the respondents agreed that it was bad. The Adere people were asked, “If your child speaks to you in Pidgin English, how does it make you feel?” All of their responses were negative—they said that they would not feel fine.

An important part of maintaining a language is to develop a writing system for it. At the time of this survey, the efforts to standardize Dzodinka has seen no progress.

6.3 Language attitudes

Adere people have a positive attitude towards the development of their mother tongue. They said they would be willing to learn to read and write in Dzodinka and have their children do the same. They are proud of their language and culture. When asked where their language was best spoken, they all said in the chief’s quarter, Bugo. They accepted that the standard dialect would be based on the language spoken in Bugo.

6.4 Development potential

John Watters (1989:376) identifies three important socioeconomic factors that have been found to affect the nature and development of a language program.

1. The homogeneity of the linguistic community (social cohesion)

2. The openness of the community to change and to better living conditions 3. The presence at the local level of a middle-aged leadership

Each of these factors as they relate to the Dzodinka-speaking communities will be discussed.

6.4.1 Homogeneity of the linguistic community (social cohesion)

Watters states, “the more homogeneous a given community is, the more chance there is for success in motivating broad based participation in the development of the language and in a mass literacy program in that community” (1989:377).

There are certain quarters that are cut off from the others during parts of the rainy season. Upper Bang reported that Nchukun, Amba, and Bangkem are cut off at some points during the year; Lower Bang reported that Upper Bang, Bangang, Nchukun, and Amba are cut off from them sometimes; Bugo reported that Amba and Nchukun are cut off from them sometimes; Bangang reported that Nchukun, Bugo, and Amba are cut off from them at times; Nchukun reported that Amba is cut off from them when there is no hanging bridge. In summary, a lack of permanent infrastructure such as bridges and properly maintained roads make travel more difficult during parts of the rainy season.

The road system between the various quarters of Adere is very limited. A four-by-four vehicle can reach Adere from Kom only in the dry season. The road network consists mainly of footpaths which handle foot traffic and now more motorcycle traffic.

All quarters reported that they are culturally and linguistically similar within each quarter and also collectively. The only exception was Amba which was reported to have a mixed population speaking both Dzodinka and Mambila. Religiously, the people of Adere share one of two religions: Christianity and African traditional religion. This acts as a uniting force for the quarters of Adere. Politically, all Adere quarters are ruled by one Fon (Chief) which creates social cohesion amongst the quarters.

6.4.2 Openness of the community to change

Quarter representatives reported that there was an organized committee for community development. The committees oversaw projects such as clearing roads, fixing hanging bridges, constructing a health clinic, repairing a secondary school, repairing a church building, cleaning paths, and maintaining latrines. The Adere people showed genuine interest in the development of their communities.

6.4.3 Presence of local middle-aged leadership

Quarter representatives reported that there was a good presence of middle-aged leadership living within each quarter. The survey team also observed this. The quarters also reported that they were confident that new leaders would take over from those currently in positions of leadership.

6.4.4 Conclusion on development potential

The potential for language development in Adere is very favorable, given the linguistic, cultural, political, and religious cohesion. There is openness to community development and a solid base of middle-aged leadership. All of these items are necessary for “widespread community participation in the development and implementation of a mass language program and for the long-term use of the language in written form” (Watters 1989:382).

6.5 Summary of results

The ALCAM wordlists elicited from Nchukun, Upper Bang, Bugo, and Bangang all showed very high levels of lexical similarity. The questionnaires administered to representatives of the quarters of Adere showed high reported comprehension among the quarters. Additionally, the findings of the

questionnaires showed a strong need and willingness on the part of the community to help develop Dzodinka. All quarters are excited about language development and accept the chief’s quarter, Bugo, as the standard speech variety. The language group is homogenous and language vitality is strong. There is also good leadership potential among middle-aged men.

7 Conclusions and recommendations

7.1 Rapid Appraisal survey of Adere quarters

One of the purposes of this report was to understand the language use patterns, vitality, attitudes, and inter-intelligibility in the quarters of Adere (Bugo, Bankem, Bangang, Nchukun, Lower Bang, and Upper Bang). The researchers sought to answer the question, “Could literature in Bugo be used in the

surrounding quarters of Adere?”

• Language use and vitality

Dzodinka is being used in all the domains of daily life. Dzodinka is the primary language used to communicate to other Adere people, whether in the market or in the fields. Pidgin English was reported to be used sometimes in these same domains. Dzodinka is also used at the health clinic and in the

churches. Since reports indicate Dzodinka is being used widely in the various domains, it is reasonable to conclude that Dzodinka is vital and viable for the long term.

• Attitudes

Traditional, government, and religious leaders in the Adere quarters are motivated to develop their mother tongue. There is social cohesion among the quarters, and they have a desire and openness to work together for development.

• Intelligibility

The results of the RA survey indicate that one standard language could be used for all of the Adere quarters. Lexicostatistics indicated that there is a very strong lexical similarity between the four main quarters of Adere. The group questionnaires indicated that the people of the Adere quarters would accept the language spoken in Bugo (the chief’s quarter) as the standard dialect.

7.2 Recorded Text Testing survey

Another purpose of this report was to describe the language intelligibility of the Koffa, Lus, and Kwaja speech varieties among neighbouring villages. The researchers sought to answer three main questions: 1. Could literature based on the Kwaja speech variety be used in the villages of Bitui, Ncha, and

Antere?

2. Could literature based on the Koffa speech variety be used in the surrounding villages of Manang, Mbat, Jui, Bang, Mballa, Kom, Mbah, and Mbepji?

3. Could literature based on the Lus speech variety be used in the village of Inkiri?

The results of the RTT survey showed that the three reference dialects of Lus, Koffa, and Kwaja can be understood by the surrounding villages. In the case of Lus, the RTT survey indicated that Inkiri has a very high comprehension level—97 percent. In the case of Koffa and the villages surrounding it, all but two villages (Mbah and Manang) scored above the 85 percent comprehension level, which suggests a high level of understanding. In the case of Kwaja, the villages of Ncha, Bitui, and Antere all showed a high comprehension level—92, 98, and 94 percent, respectively. With the exception of Mbah and Manang, the standard deviation of scores in all villages was 10 or below. This suggests that there are close linguistic similarities between each village and the reference dialect, i.e. inherent intelligibility.

Mbah and Manang were the two villages where the comprehension scores were below the 85 percent cutoff level (but above the 70 percent minimum comprehension level). The mean comprehension score for Mbah was 81.82 percent and the standard deviation was 17. The older Mbah participants (thirty-six years of age and above) scored much higher on the test than the younger ones. The standard deviations for the older and younger participants were 6 and 15, respectively. This suggests that by virtue of their age, older participants have had more time to learn the Koffa speech variety than the younger ones. If the Mbah people favor the use of Koffa materials, it is recommended that they be a part of the Koffa language project. Since Mbah is related to Koffa linguistically (comprehension scores were above 70 percent), there is a possibility that Mbah could use Koffa literature despite the standard deviation being over 15. It is recommended then, if and when Koffa is developed, testing with Koffa alphabet books and primers be done in Mbah.

In Manang the mean score was 78.9 percent and the standard deviation was 16.1. The large variation in scores suggests that those who tested well probably had considerable contact with Koffa people or even other villages close to Koffa such as Mbat or Jui. The speech variety of Manang should be considered as a distant dialect. However, it is close enough to Koffa linguistically that it is worth

including Manang in the Koffa language project. There is still a possibility that Manang could use Koffa literature, since their comprehension scores are above the 70 percent threshold. If the attitudes of the Manang people favor the use of Koffa materials, then that may overcome their possible linguistic difficulties in using Koffa.

Alternatively, if the Mbah and Manang people have difficulty in reading and writing in Koffa, further testing could be done to see if they could use literacy materials in the Limbum language.

As shown in map 3, the Mfumte language groups can be divided into four geographical regions: Northern (Adere), Southern (Koffa), Central (Lus), and Eastern (Kwaja). The names of the villages used as reference dialects for each group are enclosed in boxes. The RTT mean comprehension score for each

village is represented by a percentage figure located next to the name of the village. These percentage figures, rounded to the closest integers, refer to the reference dialect which is included in the same enclosed region.

Map 3. RTT survey summary

7.3 Recommendations for further survey

During the course of the RA survey, Adere people reported that there was a quarter called Amba located east of Bugo across the Donga River. Amba was reported to have a mixed population of Dzodinka and Mambila speakers. Unfortunately, the team was not able to reach this quarter to administer the group questionnaire and elicit a wordlist. Future research should be done to discover the lexical similarity of Amba to the reference dialect of Bugo. It would also be informative to research sociolinguistic factors such as the attitudes towards the reference dialect, social cohesion, and the openness of the community

to development. The purpose of this future research would be to determine if literature based on the Bugo speech variety could be used by people living in the quarter of Amba.

7.4 Recommendations for language development

Based on the results of the RA and RTT surveys, it is recommended that language development work be started in the Bugo speech variety of Adere, in Koffa, and in Kwaja (Lus is already under development). A precondition to language development in Bugo, Koffa, and Kwaja should be the establishment of language committees. These committees would be responsible for their respective language projects.

The first steps in developing these three varieties would be as follows:

• Collect a 1700-item wordlist, complete a phonology, and write a tentative orthography statement.

• Develop a dictionary and other literacy materials.

7.5 Recommended modifications to Ethnologue and ALCAM

Based on the RA and RTT data, the following changes are proposed for Ethnologue and ALCAM, which will involve first of all, a change to ISO 639-3.

1. Add entry for ‘Mfumte, Central’ in both ALCAM and Ethnologue. There are two villages: one in the Mfumte area northwest of Nwa, Nwa subdivision, Donga-Mantung division, North West Region and another in Taraba State, Nigeria. Niger-Congo, Atlantic-Congo, Volta-Congo, Benue-Congo, Bantoid, Southern, Wide Grassfields, Narrow Grassfields, Mbam-Nkam, Nkambe. Dialects: Inkiri. Population estimate: 6,500 (author’s estimate). Literacy rate in first language is below one percent. Literacy rate in second language is 5 to 15 percent. Christian, traditional religion.

2. Change ALCAM [904] and Ethnologue entry for ‘Dzodinka [add]’ to ‘Lidzonka’. Alternate names are Dzodinka, Adere, Northern Mfumte, Adiri, Arderi, and Dzodzinka. Keep the same three-letter Ethnologue

code [add]. Keep the same information for location and genetic classification. Population estimate: 3574. Literacy rate in first language is below one percent. Literacy rate in second language is 5 to 15 percent. Christian, traditional religion.

3. Change ALCAM [905] and Ethnologue entry for ‘Mfumte (Nfumte) [nfu]’ to ‘Mfumte, Southern’. There are nine villages in the Mfumte area northwest of Nwa and east of Nkambe, Nwa subdivision, Donga-Mantung division, North West Region. Niger-Congo, Atlantic-Congo, Volta-Congo, Benue-Congo, Bantoid, Southern, Wide Grassfields, Narrow Grassfields, Mbam-Nkam, Nkambe. Dialects: Kom, Mballa, Bang, Jui, Mbat, Manang, Mbibji (Mbepji), and Mbah. Population estimate: 12,000 (author’s estimate). Literacy rate in first language is below one percent. Literacy rate in second language is 5 to 15 percent. Christian, traditional religion.

4. Delete entry for Ndaktup [ncp].

5. Change Ethnologue entry for ‘Kwaja [kdz]’ to ‘Mfumte, Eastern’ and add ‘Mfumte, Eastern’ to ALCAM. Keep the same three-letter Ethnologue code [kdz]. Add: There are four villages in the Mfumte area north of Nwa and east of Nkambe, Nwa subdivision, Donga-Mantung division, North West Region. Ethnologue

currently says Nkambe subdivision but it should say Nwa subdivision. Niger-Congo, Atlantic-Congo, Volta-Congo, Benue-Congo, Bantoid, Southern, Wide Grassfields, Narrow Grassfields, Mbam-Nkam, Nkambe. Dialects: Bitui, Ncha, and Antere. Population estimate: 5,500 (author’s estimate). Literacy rate in first language is below one percent. Literacy rate in second language is 5 to 15 percent. Christian, traditional religion.

23

A1: Koffa RTT text

Name: Respondent 1

Village of birth: Koffa

Village where raised: Koffa

Age: 68

Sex: Male

Village of mother: Koffa

Village of father: Koffa

Number of years outside village: 2 years 9 months

Date story given: May 11, 2010

Hunting Story

(Questions and answers are in bold).

One day we were together with our children and we prayed to God.

Q1. What did they do? A. They prayed to God.

I said that I wanted to go to the bush because I am one who hunts for animals.

Q2. Where did he want to go? A. To the bush/for hunting

When we had all prayed, I stood up and picked up my gun, cartridge, bag, and cutlass and hung them on my arm.

Q3. Where did he hang his things? A. On his arm/shoulder

Then I started towards the market and after reaching and going past the market a little ways, I heard the movement of an animal (cutting grass) in the bush and I started asking myself what is was. I had doubts as to what it was, whether a mole or a cutting grass. As I took a step forward, there was a lot of noise because of the dry leaves, so I turned back.

Q4. What did he step on? A. Dry leaves (or leaves)

I came back and saw another path and carefully ascended it. I stood up and looked ahead using the light on my head which I had already turned on. I tried to find out (what it was) and I saw something like a stone but then I saw it shook itself. Then I saw that it was a cutting grass. I shot it with my gun and took and put it in my bag then I came back to the path.

Q5. Where did he put the animal? A. In his bag

I followed the path and came to a sugarcane farm, then I continued following the path to the left and came to my friend’s hut. There were palm nuts stored there.

Q6. Whose hut was it? A. His friend’s

I went behind the hut and stood there in the midst of some pineapple plants and then decided to descend from there. There was something telling me that I should come back to that place (by the hut). When I returned back to that place, I heard some movement in the bush and wondered to myself how I could get close to it. As I stood there, I listened and the thing was coming closer and closer. As I shone my light on it, I noticed that its eyes were high above the ground—it was a big animal! I shot it with my gun and it ran into the forest, groaning loudly. I removed the empty shell from the gun and quickly put another bullet inside. I thought to myself, I should not go towards it—for fear that it may be a tiger and may harm me. I left and returned to my house and asked my wife, “are you listening?”, then she said yes and I said that I had shot an animal but I did not know the type of animal because it was crying a lot. I thought it best to wait until morning before I went and looked for it. At dawn, when we finished praying, I left and as I reached the bush, I saw a dead fox.

Q7. When he reached the bush, what did he see? A. A dead fox

I carried it on my shoulder to the village and was rejoicing very much.

Q8. How did he carry the fox? A. On his shoulder/arm

I thanked God because during this one night I was able to get two animals.

Q9. Why did he thank God? A. Because he got two animals in one night

A2: Lus RTT text

Name: Respondent 2

Village of birth: Lus

Village where raised: Lus

Age: 47

Sex: Male

Village of mother: Born and raised in Lus

Village of father: Born and raised in Lus

Number of years outside of village: 19 years

Date story given: May 14, 2010

Palm Oil Story

(Questions and answers are in bold).

One day I went to the bush. I went to cut palm nuts.

Q1. He went to do what? A. To cut palm nuts

I took the climber and cutlass and climbed up the palm trees.

Q2. What did he climb up? A. Palm trees

I cut the palm nuts for a long time until there were many. I cut nuts from nine palm trees.

Q3. How many trees did he cut palm nuts from? A. Nine

On some palm trees, I cut three palm nut heads, on some two palm nut heads and on some others four palm nut heads. The palm heads were all together seventeen.

Q4. How many palm nut heads in total did he cut? A. Seventeen

I cut for a long time and was tired. I came down from the last palm tree and went and sat down. I rested a while and then started wondering what I would eat when I was hungry. When it was noon time, I thought of one mango tree that I could go to and pluck and eat. I plucked mangoes and ate for a long time and finished. I stood up and went back to my house.

Q5. Where did he go? A. To his house

I came and remained for two days (in the village). After, I went again to cut palm nuts. I went and cut palm nuts on that day but did not cut many. I cut six palm trees. I cut thirteen heads of palm nuts. The heads of palm nuts were all together thirty. I was very happy and was thinking that the time had come for me to cook palm nuts.

Q6. The Time had come for him to do what? A. Cook palm nuts.

After two days, when I had returned to the village, I went back again. I thought that I should go and prepare the nuts so that I will and cook them. I called my siblings children. There were two girls. I said that they should come so that we will pick the palm nuts and put them in the drum. That day they did not come. They failed to come. I wondered what I would do. I picked the nuts for a long time and then put them all in a basket but did not carry them. I returned back to the village. I did not carry the nuts. The next day, I still begged the girls to come. We went but only one girl came with us. The other girl did not go. We were altogether with my daughter and myself, three people.

Q7. How many people went? A. Three people

We picked the palm nuts for a long time, then we sifted and carried them to where the palm nuts were cooked at the pounding area.

Q8. Where did they carry the palm nuts to? A. Place of cooking nuts and/or the pounding area.

I looked for a drum somewhere.

Q9. What did he look for? A. Drum

I thought over and over again where will I get a drum? That day we carried the nuts and poured them

Q10. Where did they pour the palm nuts? A. On the ground

I did not know what I would do. Before I returned home, I fetched all the firewood. Then I went and looked for a drum and brought it. Reaching where I use to cook palm nuts, I collected all the palm nuts and put them into the drum and cooked them on that day. We cooked the palm nuts. After they were well cooked, we removed and pounded them for a long time and squeezed them all. We carried all the oil and put it in the drum on the day. We cooked them a long time until the oil came out. I was happy that I had oil to eat soup with.

A3: Kwaja RTT text

Name: Respondent 3

Village of birth: Kwaja

Village where raised: Kwaja

Age: 40

Sex: Male

Village of mother: Born and raised in Kwaja

Village of father: Born and raised in Kwaja

Number of years outside of village: 2 years

Date story given: May 17, 2010

Kwaja River Story

(Questions and answers are in bold) We left the village, there were eight of us.

Q1. How many people left the village? A. Eight

We travelled. When reaching at Mantung, we climbed up and reached Tep. We bought some food.

Q2. What did they buy? A. Food

When reaching the Donga river, (we saw that) it was flooded.

Q3. What did they notice about the river Donga when they reached it?

A. It was flooded.

The flood entered the maize farm.

Q4. Where did the flood enter into? A. The maize farm

When we entered the maize farm, the canoe man came with the canoe. We were eight. After we climbed inside, we went and reached the middle of the river Donga. The paddle stick broke.

Q5. What broke? A. The paddle stick

When the stick broke, I was also holding my walking stick in my hand. My own walking stick fell into the water.

Q6. Where did his walking stick go? A. Into the water

The canoe man’s paddle stick broke. All of us fell into the water. The canoe turned us all into the water. After turning us in the water, seven people drowned and were carried away by the river.

Q7. How many people drowned? A. Seven

I tried to swim and the canoe man also tried to swim. We both went to the other side of the bank. The canoe man tried to save some from being carried away by the flood, but he was unable to catch even one. He tried to catch some but he did not see any. (I came out and then) I went on my way. After I left, I reached a stream called “Eight.” I crossed and went to the other side. I climbed and reached War. After reaching War, I continued and reached Gembu and bought things.

Q8. What did he do in Gembu? A. He bought things.

I came back and reached the Donga River. I saw a big canoe there. I told the canoe man to take me across.

Q9. What did he tell the canoe man? A. To take him across

He said that I should pay him ten naira, so then I paid him.

Q10. How much money did he pay to cross back over the Donga?

A. Ten naira

I started to climb until I reached Mantung. Then I went to Ntim. Reaching Ntim, I followed through Mufeng. When I reached Mufeng, I sat down and was quiet. They asked me what happened. I was unable to talk and wanted to cry. “What happened to you?” they asked. I did not talk and wanted to cry. They

said that I should not cry because they had already heard what happened. I said, “What will I do?” One of my friends called me and said that I should drink wine.

Q11. Who told him that he should drink wine? A. One of his friends

A4: Pidgin RTT text

Name: Respondent 4

Village of birth: Lus

Village where raised: Lus (until 12 years of age)

Age: 46 (1966)

Sex: Male

Village of mother: Born and raised in Inkiri

Village of father: Born and raised in Lus

Number of years outside of village: 13 years Pidgin Goat Story

(Questions and answers are in bold.) I was there at my house.

Q1. Where was he? A. At his house

I saw a man coming. He said that his father sent him to call me for some work. I asked him what kind of work. He said that wind had removed the roof of his father’s house.

Q2. What did the wind move? A. Father’s roof

And put it on the ground. We went and reached some water called River Donga.

Q3. They came to what river? A. The Donga

We reached there and I saw the house. It was true that the wind had blown the roof down to the ground. Then I talked with the papa.

Q4. Who did he talk with? A. The papa

I said that I would do the work. I asked him how much he would pay me. He told me that he didn't have any money but he had a goat.

Q5. The papa did not have money, what did he have? A. Goat

So I started to do the work. Then I removed all the old planks.

Q6. What did he move? A. All the old planks

And put new ones there. I did this work for three days.

Q7. How many days did he work? A. Three days

I said that he should give me the goat. He gave me a small goat for that is what he said he would give me—a goat. Then I thought to myself, If I remain there too long the Donga will overflow. So, I took that

small goat.

Q8. What did he take? A. Small goat

I said that I didn’t want any trouble. I came with that goat. I wanted to reach back when one of the papa’s children (who had married some other place) said he wanted to buy that goat. He wanted to buy it for 4,000. I refused to sell it to him. I crossed the Donga River and reached Lus with that goat and sold it to a different man for that same 4,000.

No. English Upper Bang Bugo Nchukun Bangang

70 hunger (general) _{nd͡ʒe˨ nd͡ʒe˨˩ nd͡ʒe˧˨ nd͡ʒe˨˩}

71 iron (metal) _{t͡sɔk˧ t͡sɔk˧ t͡sɔk˧ t͡sɔk˧}

72 one m_ɔ˨ʃin˧ m_ɔ˧˨ m_ɔ˨ʃin˧ m_ɔ˨˧˨

73 two _{ba˦˧ ba˦˧ ba˦˧ ba˨˧˨}

74 three _{t͡ʃɛt˨˦ t͡ʃet˨˦ t͡ʃet˨˦ t͡ʃet˨˦}

75 four _{ko˦˨ ko˦˨ ko˦˨ ko˦˨}

76 five tan˧˨ tan˧˨ tan˧˨ tan˧˨

77 six to:˨va˧ to:˨va˧ to:˨va˧ nto:˨va˧

78 seven za˨mbɛ˨˩ za˨mbɛ˨˩ za˨mbɛ˨˩ za˨mbɛ˨˩

79 eight wa:˨mɔ˧ wa:˨mɔ˧ wa:˨mɔ˧ wa:˨mɔ˧

80 nine bu:˨kwɔ˧˨ bu:˨kwɔ˧˨ bu:˨kwɔ˧˨ bu:˨kwɔ˧˨

81 ten wum˧˨ wum˨˩ wum˨˩ wum˨˩

82 come wɛ˧˨ wɛ˧˨ wɛ˧˨ wɛ˧˨

83 send (someone) to˦nd͡za˦˨ tu˦nd͡za˦˨ tu˦nd͡za˦˨ tu˦nda˦˨

84 walk d͡ʒi˧nɛ˨ d͡ʒi˧nɛ˨˩ d͡ʒi˧nɛ˨˩ d͡ʒi˧nə˨˩

85 fall ŋɡwɔŋ˨ gwɔ˧˨ gwɔ˧˨ gwɔ˧˨

86 leave lɔ˨˩ lɔ˧˨ lɔ˧˨ lɔ˧˨

87 fly li˦˨ li˦˨ li˦˨ li˦˨

88 pour _{ku˨tə˧˨ ku˨tə˧˨ ku˨tə˧˨ ku˨tə˧˨}

89 strike _{li˧vɛ˨˩ li˧bɛ˨˩ li˧bɛ˨˩ lip˧˨}

90 bite _{lo˦mɛ˨˩ lo˦mɛ˨˩ lom˧ lo˦mɛ˨˩}

91 wash (transitive) _{so˨kə˧˨ so˨kə˧˨ so˨kən˧˨ so˨kə˧˨}

92 split (wood) _{sa˦nə˦˨ sa˦nəŋ˦˨ sa˦nə˦˨ sa˦nə˦˨}

93 give f_ɛ˦˨ f_ɛ˦˨ f_ɛn˦˨ f_ɛ˦˨

94 steal _{ɣa˧rɨ˧˨ ɣa˧rɨ˧˨ ɣa˧lɨn˧˨ wu˧ɣə˧˨}

95 squeeze ka˦mɛ˨ ka˦mɛ˨˩ ka˦mɛ˨ ka˦mɛ˨˩

96 cultivate _{fa˦˨ fa˦ŋɛ˨ fa˦ŋɛ˨ fa˦ŋɛ˨˩}

97 bury (transitive) t_ɔ˦ɲɨn˨ nt_ɔ˦ŋin˨ t_ɔ˦ŋi˨ t_ɔ˦ŋi˧˨

98 burn (transitive) tɔ˦nɛ˨ tɔ˦nɛ˨˩ tɔ˦nɛ˨˩ tɔ˦nɛ˧˨

99 eat le˦˨ le˦˨ le˦˨ le˦˨

100 drink nɔ˦˨ nɔ˦˨ nɔ˦˨ nɔ˦˨

101 vomit t͡ʃɛ˦kə˦˨ t͡ʃwe˦kə˦˨ t͡ʃwe˦kə˦˨ t͡ʃwe˦kə˦˨

102 suck ɲɔ˦˨ ɲɔ˦˨ ɲɔ˦˨ ɲɔ˦˨

103 spit (saliva) t͡ʃwe˦˨ t͡ʃwe˦˨ t͡ʃwe˦˨ t͡ʃwe˦˨

104 blow (on) fə˦mi˨˩ fə˦mi˨˩ fə˦mi˨˩ fə˦mi˨˩

45

No. English Upper Bang Bugo Nchukun Bangang

106 give birth _{mbin˦ bi˦˨ bi˦˨ bi˦˨}

107 die _{li˦kwɔ˦ li˨kwɔ˧} kw_{ɔ˧˨ li˨kwɔ˧}

108 kill _{jun˦ jun˦ ju˦˨ jun˦}

109 push _t͡sɨ˧n_{ɛ˨ t͡sɨ˧}n_{ɛ˨˩ mu˧˨ t͡sɨ˧}n_ɛ˨˩

110 pull _{sun˦ su˦ŋɛ˨ su˦˨ sun˦}

111 sing _ɣə˦m_{ɛ˨ ɣə˦}m_{ɛ˨ ɣə˦}m_{ɛ˧˨ ɣə˦}m_ɛ˨

112 play (a game) mɔ˦ŋɛ˨ li˨mɔ˨˩ li˨mɔ˨˩ mɔ˦ŋɛ˨

113 be afraid wɔ˦vɛ˨ wɔ˦bɛ˧ wɔ˦bɛn˧ wɔ˦bɛ˧

114 want kɔ˧ŋɛ˨ kɔ˧ŋɛ˨ kɔ˧mun˧˨ kɔ˧ŋɛ˨

115 say ta˦ndi˧˨ ta˦ndi˧˨ ta˦ndi˧˨ ta˦ndi˧˨

116 see jɛ˦ŋə˨ je˧˨ je˧ jɛ˧

117 show lɨ˧nə˨ lɨ˨nən˨˩ lɨ˨nən˨˩ lɨ˨nən˨˩

118 hear wu˦ŋɛ˨ wun˦ wən˦ wun˦

119 know li˧ŋɛ˨ lin˨ lin˨ lin˨

120 count ta˦ŋə˨ ntan˧ tan˧ ntan˧

Researchers: Novester Nweshie and Greg McLean Upper Bang wordlist

Informant: male, age 48 Date: June 30, 2010 Location: Lus

Bugo wordlist

Informant: male, age 25 Date: July 6, 2010 Location: Lus

Nchukun wordlist Informant: male, age 50 Date: June 30, 2010 Location: Lus

Bangang wordlist

Informant: male, age 60 Date: June 29, 2010 Location: Lus Notes:

(1) When the author elicited the verbs (see items 82–120), most were in their present tense form but in some rare cases, the imperative form was taken accidentally. The imperative form of the verb has the suffix ŋɛ/ŋə (e.g. 114, 116, 119, and 120). In item 116 for example, the word ‘to see’ is jɛŋə in Upper Bang and je or jɛ in the other three quarters. The difference in these verb forms therefore is grammatical (i.e. not lexical).

(2) The word for ‘claw’ _{(12) is ŋɡak}, _{not ŋɡabɔ}. _{Ŋɡabɔ means ‘fingernail’}. Therefore, it is not clear if there is a lexical difference between the four quarters.

(3) There is no indigenous word for ‘tortoise’ (59) in the four quarters—only the pidgin word ‘toroki’. (4) The Nchukun man’s mother and father are from Bangang and Bankem, respectively. Nchukun is a relatively new settlement. Nchukun consists of about ten compounds, with perhaps a total of thirty to seventy people.

Appendix E: Religious Leader Questionnaire

49

References

Baeke, Viviane. 2004. Le temps des rites. Ordre du monde et destin individuel en pays Wuli (Cameroun). Nanterre: Societe d’Ethnologie.

Bergman, Ted G. 1991. Rapid appraisal of languages. Notes on Scripture in Use and Language Programs 28:3–11. Dallas: Summer Institute of Linguistics.

Casad, Eugene H. 1974. Dialect intelligibility testing. Dallas: Summer Institute of Linguistics.

Dieu, M., and Renaud, P. eds. 1983. Situation linguistique en Afrique Centrale: Inventaire préliminaire, le Cameroun. In Atlas Linguistique de l’Afrique Centrale (ALAC), Atlas Linguistique du Cameroun (ALCAM). Paris/ Yaoundé: ACCT; DRGST; CERDOTOLA.

Eyoh, Julius A. 2008. Mfumte phonology and revitalisation of the language. D.E.A. dissertation. Yaoundé, Cameroon: University of Yaounde.

Eyoh, Julius A., and Robert Hedinger. 2008. Mfumte orthography guide. Yaounde, Cameroon: The National Association of Cameroonian Language Committees.

http://www.silcam.org/languages/languagepage.php?languageid=214, accessed June 24, 2014.

Grant, Caroline A., Margaret A. Griffin, Joseph Mbongue, and Lawrence M. Seguin. 1994. A Rapid Appraisal survey of Dzodinka (ALCAM 904) and Mfumte (ALCAM 905), Donga Mantung division, Northwest Province. Yaoundé, Cameroon: Summer Institute of Linguistics.

http://www.silcam.org/languages/languagepage.php?languageid=3, accessed June 24, 2014.

Grimes, Barbara F. 1987. How bilingual is bilingual? Notes on Linguistics 40A:34–54 Dallas: Summer Institute of Linguistics.

Grimes, Joseph E. 1995. Language survey reference guide. Dallas: Summer Institute of Linguistics. Lewis, M. Paul, Gary F. Simons, and Charles D. Fennig (eds.). 2014. Ethnologue: Languages of the world.

Seventeenth edition. Dallas: SIL International. http://www.ethnologue.com, accessed June 24, 2014. Stalder, Jürg. 1996a. Rapid appraisal. Notes on Literacy in Use and Language Programs 48(6):5–23. Dallas:

Summer Institute of Linguistics.

Stalder, Jürg. 1996b. Update on rapid appraisal. Notes on Literacy in Use and Language Programs 48(6):24–28. Dallas: Summer Institute of Linguistics.

Voorhoeve, Jan. 1980. Noun classes in Adere. In Larry Hyman (ed.), Noun Classes in the Grassfields Bantu Borderland. Southern California Occasional Papers in Linguistics 8:57–72. Los Angeles: University of Southern California.

Watters, John R. 1989. Three socio-economic factors affecting the nature and development of language programs. In Ted G. Bergman (comp.), Survey reference manual, 6.7.1–6.7.12. Dallas: Summer Institute of Linguistics.